Humans Wanted: Do Not Leave Information Curation to AI
March 20, 2024
This essay is the work of a dumb dinobaby. No smart software required.
Remember RSS feeds? Before social media took over the Internet, they were the way we got updates from sources we followed. It may be time to dust off the RSS, for it is part of blogger Joan Westenberg’s plan to bring a human touch back to the Web. We learn of her suggestions in, “Curation Is the Last Best Hope of Intelligent Discourse.”
Westenberg argues human judgement is essential in a world dominated by AI-generated content of dubious quality and veracity. Generative AI is simply not up to the task. Not now, perhaps not ever. Fortunately, a remedy is already being pursued, and Westenberg implores us all to join in. She writes:
“Across the Fediverse and beyond, respected voices are leveraging platforms like Mastodon and their websites to share personally vetted links, analysis, and creations following the POSSE model – Publish on your Own Site, Syndicate Elsewhere. By passing high-quality, human-centric content through their own lens of discernment before syndicating it to social networks, these curators create islands of sanity amidst oceans of machine-generated content of questionable provenance. Their followers, in turn, further syndicate these nuggets of insight across the social web, providing an alternative to centralised, algorithmically boosted feeds. This distributed, decentralised model follows the architecture of the web itself – networks within networks, sites linking out to others based on trust and perceived authority. It’s a rethinking of information democracy around engaged participation and critical thinking from readers, not just content generation alone from so-called ‘influencers’ boosted by profit-driven behemoths. We are all responsible for carefully stewarding our attention and the content we amplify via shares and recommendations. With more voices comes more noise – but also more opportunity to find signals of truth if we empower discernment. This POSSE model interfaces beautifully with RSS, enabling subscribers to follow websites, blogs and podcasts they trust via open standard feeds completely uncensored by any central platform.”
But is AI all bad? No, Westenberg admits, the technology can be harnessed for good. She points to Anthropic‘s Constitutional AI as an example: it was designed to preserve existing texts instead of overwriting them with automated content. It is also possible, she notes, to develop AI systems that assist human curators instead of compete with them. But we suspect we cannot rely on companies that profit from the proliferation of shoddy AI content to supply such systems. Who will? People with English majors?
Cynthia Murrell, March 20, 2024
Software Failure: Why Problems Abound and Multiply Like Gerbils
March 19, 2024
This essay is the work of a dumb dinobaby. No smart software required.
I read “Why Software Projects Fail” after a lunch at which crappy software and lousy products were a source of amusement. The door fell off what?
What’s interesting about the article is that it contains a number of statements which resonated with me. I recommend the article, but I want to highlight several statements from the essay. These do a good job of explaining why small and large projects go off the rails. Within the last 12 months I witnessed one project get tangled in solving a problem that existed 15 years ago. Today not so much. The team crafted the equivalent of a Greek Corinthian helmet from the 8th century BCE. Another project infused with AI and vision of providing a “new” approach to security wobble between and among a telecommunications approach, an email approach, and an SMS approach with bells and whistles only a science fiction fan would appreciate. Both of these examples obtained funding; neither set out to build a clown car. What happened? That’s where “Why Projects Fail?” becomes relevant.
Thanks, MSFT Copilot. You have that MVP idea nailed with the recent Windows 11 update, don’t you. Good enough I suppose.
Let’s look at three passages from the essay, shall we?
Belief in One’s Abilities or I Got an Also-Participated Ribbon in Middle School
Here’s the statement from the essay:
One of the things that I’ve noticed is that developers often underestimate not just the complexity of tasks, but there’s a general overconfidence in their abilities, not limited by programming:
- Overconfidence in their coding skills.
- Overconfidence in learning new technologies.
- Overconfidence in our abstractions.
- Overconfidence in external dependencies, e.g., third-party services or some open-source library.
My comment: Spot on. Those ribbons built confidence, but they mean nothing.
Open Source Is Great Unless It Has Been Screwed Up, Become a Malware Delivery Vehicle, or Just Does Not Work
Here’s the statement from the essay:
… anything you do not directly control is a risk of hidden complexity. The assumption that third-party services, libraries, packages, or APIs will work as expected without bugs is a common oversight.
My view is that “complexity” is kicked around as if everyone held a shared understanding of the term. There are quite different types of complexity. For software, there is the complexity of a simple process created in Assembler but essentially impenetrable to a 20-something from a whiz-bang computer science school. There is the complexity of software built over time by attention deficit driven people who do not communicate, coordinate, or care what others are doing, will do, or have done. Toss in the complexity of indifferent, uninformed, or uninterested “management,” and you get an exciting environment in which to “fix up” software. The cherry on top of this confection is that quite a bit of software is assumed to be good. Ho ho ho.
The Real World: It Exists and Permeates
I liked this statement:
Technology that seemed straightforward refuses to cooperate, external competitors launch similar ideas, key partners back out, and internal business stakeholders focus more on the projects that include AI in their name. Things slow down, and as months turn into years, enthusiasm wanes. Then the snowball continues — key members leave, and new people join, each departure a slight shift in direction. New tech lead steps in, eager to leave their mark, steering the project further from its original course. At this point, nobody knows where the project is headed, and nobody wants to admit the project has failed. It’s a tough spot, especially when everyone’s playing it safe, avoiding the embarrassment or penalties of admitting failure.
What are the signals that trouble looms? A fumbled ball at the Google or the Apple car that isn’t can be blinking lights. Staff who go rogue on social media or find an ambulance chasing honed law firm can catch some individual’s attention.
The write up contains other helpful observations. Will people take heed? Are you kidding me? Excellence costs money, requires informed judgment, and expertise. Who has time for this with AI calendars, the demands of TikTok and Instagram, and hitting the local coffee shop?
Stephen E Arnold, March 19, 2024
A Single Google Gem for March 19, 2024
March 19, 2024
This essay is the work of a dumb dinobaby. No smart software required.
I want to focus on what could be the star sapphire of Googledom. The story appeared on the estimable Murdoch confection Fox News. Its title? “Is Google Too Broken to Be Fixed? Investors Deeply Frustrated and Angry, Former Insider Warns”? The word choice in this Googley headline signals the alert reader that the Foxy folks have a juicy story to share. “Broken,” “Frustrated,” “Angry,” and “Warns” suggest that someone has identified some issues at the beloved Google.
A Google gem. Thanks, MSFT Copilot Bing thing. How’s the staff’s security today?
The write up states:
A former Google executive [David Friedberg] revealed that investors are “deeply frustrated” that the scandal surrounding their Gemini artificial intelligence (AI) model is becoming a “real threat” to the tech company. Google has issued several apologies for Gemini after critics slammed the AI for creating “woke” content.
The Xoogler, in what seems to be tortured prose, allegedly said:
“The real threat to Google is more so, are they in a position to maintain their search monopoly or maintain the chunk of profits that drive the business under the threat of AI? Are they adapting? And less so about the anger around woke and DEI,” Friedberg explained. “Because most of the investors I spoke with aren’t angry about the woke, DEI search engine, they’re angry about the fact that such a blunder happened and that it indicates that Google may not be able to compete effectively and isn’t organized to compete effectively just from a consumer competitiveness perspective,” he continued.
The interesting comment in the write up (which is recycled podcast chatter) seems to be:
Google CEO Sundar Pichai promised the company was working “around the clock” to fix the AI model, calling the images generated “biased” and “completely unacceptable.”
Does the comment attributed to a Big Dog Microsoftie reflect the new perception of the Google. The Hindustan Times, which should have radar tuned to the actions, of certain executives with roots entwined in India reported:
Satya Nadella said that Google “should have been the default winner” of Big Tech’s AI race as the resources available to it are the maximum which would easily make it a frontrunner.
My interpretation of this statement is that Google had a chance to own the AI casino, roulette wheel, and the croupiers. Instead, Google’s senior management ran over the smart squirrel with the Paris demonstration of the fantastic Bard AI system, a series of me-too announcements, and the outputting of US historical scenes with people of color turning up in what I would call surprising places.
Then the PR parade of Google wizards explains the online advertising firm’s innovations in playing games, figuring out health stuff (shades of IBM Watson), and achieving quantum supremacy in everything. Well, everything except smart software. The predicament of the ad giant is illuminated with the burning of billions in market cap coincident with the wizards’ flubs.
Net net: That’s a gem. Google losing a game it allegedly owned. I am waiting for the next podcast about the Sundar & Prabhakar Comedy Tour.
Stephen E Arnold, March 19, 2024
Microsoft Decides to Work with CISPE on Cloudy Concerns
March 19, 2024
This essay is the work of a dumb dinobaby. No smart software required.
Perhaps a billion and a half dollars in fines can make a difference to a big tech company after all. In what looks like a move to avoid more regulatory scrutiny, Yahoo Finance reports, “Microsoft in Talks to End Trade Body’s Cloud Computing Complaint.” The trade body here is CISPE, a group of firms that provide cloud services in Europe. Amazon is one of those, but 26 smaller companies are also members. The group asserts certain changes Microsoft made to its terms of service in October of 2022 have harmed Europe’s cloud computing ecosystem. How, exactly, is unclear. Writer Foo Yun Chee tells us:
“[CISPE] said it had received several complaints about Microsoft, including in relation to its product Azure, which it was assessing based on its standard procedures, but declined to comment further. Azure is Microsoft’s cloud computing platform. CISPE said the discussions were at an early stage and it was uncertain whether these would result in effective remedies but said ‘substantive progress must be achieved in the first quarter of 2024’. ‘We are supportive of a fast and effective resolution to these harms but reiterate that it is Microsoft which must end its unfair software licensing practices to deliver this outcome,’ said CISPE secretary general Francisco Mingorance. Microsoft, which notched up 1.6 billion euros ($1.7 billion) in EU antitrust fines in the previous decade, has in recent years changed its approach towards regulators to a more accommodative one.”
Just how accommodating with Microsoft will be remains to be seen.
Cynthia Murrell, March 19, 2024
Old Code, New Code: Can You Make It Work Again… Sort Of?
March 18, 2024
This essay is the work of a dumb dinobaby. No smart software required.
Even hippy dippy super slick AI start ups have a technical debt problem. It is, in my opinion, no different from the “costs” imposed on outfits like JPMorgan Chase or (heaven help us) AMTRAK. Software which mostly works is subject to two environmental problems. First, the people who wrote the code or made it work that last time catastrophe struck (hello, AT&T, how are those pushed updates working for you now?) move on, quit, or whatever. Second, the technical options for remediating the problem are evolving (how are those security hot fixes working out, Microsoft?).
The helpful father asks an question the aspiring engineer cannot answer. Thus it was when the wizard was a child, and it is when the wizard is working on a modern engineering project. Buildings tip; aircraft lose doors and wheels. Software updates kill computers. Self-driving cars cannot. Thanks, MSFT Copilot. Did you get your model airplane to fly when you were a wee lad? I think I know the answer.
I thought about this problem of the cost of code remediating, fixing, redoing, upgrading or whatever term fast-talking sales engineers use in their Zooms and PowerPoints as I read “The High-Risk Refactoring.” The write up does a good job of explaining in a gentle way what happens when suits authorize making old code like new again. (The suits do not know the agonies of the original developers, but why should “history” intrude on a whiz bang GenX or GenY management type?
The article says:
it’s highly important to ensure the system works the same way after the swap with the new code. In that regard, immediately spotting when something breaks throughout the whole refactoring process is very helpful. No one wants to find that out in production.
No kidding.
In most cases, there are insufficient skilled people and money to create a new or revamped system, get it up and running in parallel for an appropriate period of time, identify the problems, remediate them, and then make the cut over. People buy cars this way, but that’s not how most organizations, regardless of size, “do” software. Okay, the take your car in, buy a new one, and drive off will not work in today’s business environment.
The write up focuses on what most organizations do; that is, write or fix new code and stick it into a system. There may or may not be resources for a staging server, but the result is the same. The old software has been “fixed” and the documentation is “sort of written” and people move on to other work or in the case of consulting engineering firms, just get replaced by a new, higher margin professional.
The write up takes a different approach and concludes with four suggestions or questions to ask. I quote:
“Refactor if things are getting too complicated, but stop if can’t prove it works.
Accompany new features with refactoring for areas you foresee to be subject to a change, but copy-pasting is ok until patterns arise.
Be proactive in finding new ways to ensure refactoring predictability, but be conservative about the assumption QA will find all the bugs.
Move business logic out of busy components, but be brave enough to keep the legacy code intact if the only argument is “this code looks wrong”.
These are useful points. I would like to suggest some bright white lines for those who have to tackle an IRS-mainframe- or AT&T-billing system type of challenge as well as tweaking an artificial intelligence solution to respond to those wonky multi-ethnic images Google generated in order to allow the Sundar & Prabhakar Comedy Team to smile sheepishly and apologize again for lousy software.
Are you ready? Let’s go:
- Fixes add to the complexity of the code base. As time goes stumbling forward, the complexity of the software becomes greater. The cost of making sure the fix works and does not create exciting dependency behavior goes up. Thus, small fixes “cost” more, and these costs are tough to control.
- The safest fixes are “wrappers”; that is, no one in his or her right mind wants to change software written in 1978 for a machine no longer in production by the manufacturer. Therefore, new software is written to interact in a “safe” way with the original software. The new code “fixes up” the problem without screwing up what grandpa programmer wrote almost half a century ago. The problem is that “wrappers” tend to slow stuff down. The fix is to say one will optimize the system while one looks for a new project or job.
- The software used for “fixing” a problem is becoming the equivalent of repairing an aircraft component with Dawn laundry detergent. The “fix” is cheap, easy to use, and good enough. The software equivalent of this Dawn solution is that it will not stand the test of time. Instead of code crafted in good old COBOL or Assembler, we have some Fancy Dan tools which may fall out of favor in a matter of months, not decades.
Many projects result in better, faster, and cheaper. The reminder “Pick two” is helpful.
Net net: Fixing up lousy or flawed software is going to increase risks and costs. The question asked by bean counters is, “How much?” The answer is, “No one knows until the project is done … if ever.”
Stephen E Arnold, March 18, 2024
Worried about TikTok? Do Not Overlook CapCut
March 18, 2024
This essay is the work of a dumb dinobaby. No smart software required.
I find the excitement about TikTok interesting. The US wants to play the reciprocity card; that is, China disallows US apps so the US can ban TikTok. How influential is TikTok? US elected officials learned first hand that TikTok users can get messages through to what is often a quite unresponsive cluster of elected officials. But let’s leave TikTok aside.
Thanks, MSFT Copilot. Good enough.
What do you know about the ByteDance cloud software CapCut? Ah, you have never heard of it. That’s not surprising because it is aimed at those who make videos for TikTok (big surprise) and other video platforms like YouTube.
CapCut has been gaining supporters like the happy-go-lucky people who published “how to” videos about CapCut on YouTube. On TikTok, CapCut short form videos have tallied billions of views. What makes it interesting to me is that it wants to phone home, store content in the “cloud”, and provide high-end tools to handle some tricky video situations like weird backgrounds on AI generated videos.
The product CapCut was named (I believe) JianYing or Viamaker (the story varies by source) which means nothing to me. The Google suggests its meanings could range from hard to paper cut out. I am not sure I buy these suggestions because Chinese is a linguistic slippery fish. Is that a question or a horse? In 2020, the app got a bit of shove into the world outside of the estimable Middle Kingdom.
Why is this important to me? Here are my reasons for creating this short post:
- Based on my tests of the app, it has some of the same data hoovering functions of TikTok
- The data of images and information about the users provides another source of potentially high value information to those with access to the information
- Data from “casual” videos might be quite useful when the person making the video has landed a job in a US national laboratory or in one the high-tech playgrounds in Silicon Valley. Am I suggesting blackmail? Of course not, but a release of certain imagery might be an interesting test of the videographer’s self-esteem.
If you want to know more about CapCut, try these links:
- Download (ideally to a burner phone or a PC specifically set up to test interesting software) at www.capcut.com
- Read about the company CapCut in this 2023 Recorded Future write up
- Learn about CapCut’s privacy issues in this Bloomberg story.
Net net: Clever stuff but who is paying attention. Parents? Regulators? Chinese intelligence operatives?
Stephen E Arnold, March 18, 2024
AI in Action: Price Fixing Play
March 18, 2024
This essay is the work of a dumb dinobaby. No smart software required.
If a tree falls in a forest and no one is there to hear it, does I make a sound? The obvious answer is yes, but the philosophers point out how can that be possible if there wasn’t anyone there to witness the event. The same argument can be made that price fixing isn’t illegal if it’s done by an AI algorithm. Smart people know it is a straw man’s fallacy and so does the Federal Trade Commission: “Price Fixing By Algorithm Is Still Price Fixing.”
The FTC and the Department of Justice agree that if an action is illegal for a human then it is illegal for an algorithm too. The official nomenclature is antitrust compliance. Both departments want to protect consumers against algorithmic collision, particularly in the housing market. They failed a joint legal brief that stresses the importance of a fair, competitive market. The brief stated that algorithms can’t be used to evade illegal price fixing agreements and it is still unlawful to share price fixing information even if the conspirators retain pricing discretion or cheat on the agreement.
Protecting consumers from unfair pricing practices is extremely important as inflation has soared. Rent has increased by 20% since 2020, especially for lower-income people. Nearly half of renters also pay more than 30% of their income in rent and utilities. The Department of Justice and the FTC also hold other industries accountable for using algorithms illegally:
“The housing industry isn’t alone in using potentially illegal collusive algorithms. The Department of Justice has previously secured a guilty plea related to the use of pricing algorithms to fix prices in online resales, and has an ongoing case against sharing of price-related and other sensitive information among meat processing competitors. Other private cases have been recently brought against hotels(link is external) and casinos(link is external).”
Hopefully the FTC and the Department of Justice retain their power to protect consumers. Inflation will continue to rise and consumers continue to suffer.
Whitney Grace, March 18, 2024
Harvard University: William James Continues Spinning in His Grave
March 15, 2024
This essay is the work of a dumb dinobaby. No smart software required.
William James, the brother of a novelist which caused my mind to wander just thinking about any one of his 20 novels, loved Harvard University. In a speech at Stanford University, he admitted his untoward affection. If one wanders by William’s grave in Cambridge Cemetery (daylight only, please), one can hear a sound similar to a giant sawmill blade emanating from the a modest tombstone. “What’s that horrific sound?” a by passer might ask. The answer: “William is spinning in his grave. It a bit like a perpetual motion machine now,” one elderly person says. “And it is getting louder.”
William is spinning in his grave because his beloved Harvard appears to foster making stuff up. Thanks, MSFT Copilot. Working on security today or just getting printers to work?
William is amping up his RPMs. Another distinguished Harvard expert, professor, shaper of the minds of young men and women and thems has been caught fabricating data. This is not the overt synthetic data shop at Stanford University’s Artificial Intelligence Lab and the commercial outfit Snorkel. Nope. This is just a faculty member who, by golly, wanted to be respected it seems.
The Chronicle of Higher Education (the immensely popular online information service consumed by thumb typers and swipers) published “Here’s the Unsealed Report Showing How Harvard Concluded That a Dishonesty Expert Committed Misconduct.” (Registration required because, you know, information about education is sensitive and users must be monitored.) The report allegedly required 1,300 pages. I did not read it. I get the drift: Another esteemed scholar just made stuff up. In my lingo, the individual shaped reality to support her / its vision of self. Reality was not delivering honor, praise, rewards, money, and freedom from teaching horrific undergraduate classes. Why not take the Excel macro to achievement: Invent and massage information. Who is going to know?
The write up says:
the committee wrote that “she does not provide any evidence of [research assistant] error that we find persuasive in explaining the major anomalies and discrepancies.” Over all, the committee determined “by a preponderance of the evidence” that Gino “significantly departed from accepted practices of the relevant research community and committed research misconduct intentionally, knowingly, or recklessly” for five alleged instances of misconduct across the four papers. The committee’s findings were unanimous, except for in one instance. For the 2012 paper about signing a form at the top, Gino was alleged to have falsified or fabricated the results for one study by removing or altering descriptions of the study procedures from drafts of the manuscript submitted for publication, thus misrepresenting the procedures in the final version. Gino acknowledged that there could have been an honest error on her part. One committee member felt that the “burden of proof” was not met while the two other members believed that research misconduct had, in fact, been committed.
Hey, William, let’s hook you up to a power test dynamometer so we can determine exactly how fast you are spinning in your chill, dank abode. Of course, if the data don’t reveal high-RPM spinning, someone at Harvard can be enlisted to touch up the data. Everyone seems to be doing from my vantage point in rural Kentucky.
Is there a way to harness the energy of professors who may cut corners and respected but deceased scholars to do something constructive? Oh, look. There’s a protest group. Let’s go ask them for some ideas. On second thought… let’s not.
Stephen E Arnold, March 15, 2024
Stanford: Tech Reinventing Higher Education: I Would Hope So
March 15, 2024
This essay is the work of a dumb dinobaby. No smart software required.
I read “How Technology Is Reinventing Education.” Essays like this one are quite amusing. The ideas flow without important context. Let’s look at this passage:
“Technology is a game-changer for education – it offers the prospect of universal access to high-quality learning experiences, and it creates fundamentally new ways of teaching,” said Dan Schwartz, dean of Stanford Graduate School of Education (GSE), who is also a professor of educational technology at the GSE and faculty director of the Stanford Accelerator for Learning. “But there are a lot of ways we teach that aren’t great, and a big fear with AI in particular is that we just get more efficient at teaching badly. This is a moment to pay attention, to do things differently.”
I
A university expert explains to a rapt audience that technology will make them healthy, wealthy, and wise. Well, that’s the what the marketing copy which the lecturer recites. Thanks, MSFT Copilot. Are you security safe today? Oh, that’s too bad.
I would suggest that Stanford’s Graduate School of Education consider these probably unimportant points:
- The president of Stanford University resigned allegedly because he fudged some data in peer-reviewed documents. True or false. Does it matter? The fellow quit.
- The Stanford Artificial Intelligence Lab or SAIL innovated with cooking up synthetic data. Not only was synthetic data the fast food of those looking for cheap and easy AI training data, Stanford became super glued to the fake data movement which may be good or it may be bad. Hallucinating is easier if the models are training using fake information perhaps?
- Stanford University produced some outstanding leaders in the high technology “space.” The contributions of famous graduates have delivered social media, shaped advertising systems, and interesting intelware companies which dabble in warfighting and saving lives from one versatile software and consulting platform.
The essay operates in smarter-than-you territory. It presents a view of the world which seems to be at odds with research results which are not reproducible, ethics-free researchers, and an awareness of how silly it looks to someone in rural Kentucky to have a president accused of pulling a grade-school essay cheating trick.
Enough pontification. How about some progress in remediating certain interesting consequences of Stanford faculty and graduates innovations?
Stephen E Arnold, March 15, 2024
Humans Wanted: Do Not Leave Information Curation to AI
March 15, 2024
This essay is the work of a dumb dinobaby. No smart software required.
Remember RSS feeds? Before social media took over the Internet, they were the way we got updates from sources we followed. It may be time to dust off the RSS, for it is part of blogger Joan Westenberg’s plan to bring a human touch back to the Web. We learn of her suggestions in, “Curation Is the Last Best Hope of Intelligent Discourse.”
Westenberg argues human judgement is essential in a world dominated by AI-generated content of dubious quality and veracity. Generative AI is simply not up to the task. Not now, perhaps not ever. Fortunately, a remedy is already being pursued, and Westenberg implores us all to join in. She writes:
“Across the Fediverse and beyond, respected voices are leveraging platforms like Mastodon and their websites to share personally vetted links, analysis, and creations following the POSSE model – Publish on your Own Site, Syndicate Elsewhere. By passing high-quality, human-centric content through their own lens of discernment before syndicating it to social networks, these curators create islands of sanity amidst oceans of machine-generated content of questionable provenance. Their followers, in turn, further syndicate these nuggets of insight across the social web, providing an alternative to centralised, algorithmically boosted feeds. This distributed, decentralised model follows the architecture of the web itself – networks within networks, sites linking out to others based on trust and perceived authority. It’s a rethinking of information democracy around engaged participation and critical thinking from readers, not just content generation alone from so-called ‘influencers’ boosted by profit-driven behemoths. We are all responsible for carefully stewarding our attention and the content we amplify via shares and recommendations. With more voices comes more noise – but also more opportunity to find signals of truth if we empower discernment. This POSSE model interfaces beautifully with RSS, enabling subscribers to follow websites, blogs and podcasts they trust via open standard feeds completely uncensored by any central platform.”
But is AI all bad? No, Westenberg admits, the technology can be harnessed for good. She points to Anthropic‘s Constitutional AI as an example: it was designed to preserve existing texts instead of overwriting them with automated content. It is also possible, she notes, to develop AI systems that assist human curators instead of compete with them. But we suspect we cannot rely on companies that profit from the proliferation of shoddy AI content to supply such systems. Who will?
Cynthia Murrell, March 15, 2024
-
- Subscribe to Beyond Search
Feature archive
News archive
-
